OPTIMIZATION OF PARTICLE IMAGE VELOCIMETERS .2. MULTIPLE PULSED SYSTEMS

The spatial resolution, detection rate and reliability of a particle image velocimeter depend critically upon careful selection of a number of parameters of the PIV system and the fluid motion. From earlier work, an analytic model and a Monte Carlo computer simulation have been generalized to analyse the effects of experimental parameters and to optimize system parameters in a multiple pulsed PIV system. The set of six non-dimensional parameters that were most significant in optimizing double pulsed PIV performance are generalized for a multiple pulsed system and shown to be similarly significant. They are the data validation criterion, the particle image density, the relative in-plane image displacement, the relative out-of-plane displacement, the velocity gradient and the ratio of the mean image displacement to the interrogation spot diameter. It is shown that a single transformation can be applied to the method of interrogation by autocorrelation analysis to extend these results to interrogation by two dimensional Fourier transform analysis of the Young's fringes. Thus, the non-dimensional parameters are studied for the case of interrogation by autocorrelation analysis. It is shown that optimal parameters for a multiple pulsed system can be generalised from earlier results for a double pulsed system. A multiple pulsed system performs optimally when the density of image pairs exceeds 10-20 and the maximum relative in-plane and out-of-plane displacements are limited to 25%. Velocity gradients reduce the signal strength causing a loss of valid data, and produce statistical bias. Expressions for the statistical bias of three dimensional velocity fields are developed and corrections for this bias are presented. Furthermore, recommendations for experimental procedures to minimize the bias effects and reduction of signal strength are outlined.